A nonlinear biomechanical model based registration method for aligning prone and supine MR breast images

Area of Science:

• Medical Imaging
• Biomechanical Modeling
• Surgical Planning

Background:

• Preoperative MR breast images offer detailed 3D information for cancer surgery.
• Aligning prone MR images with supine surgical positioning is challenging due to significant breast deformation.
• Existing methods like nonrigid registration or biomechanical modeling alone are insufficient for large deformations.

Purpose of the Study:

• To develop and evaluate a nonlinear biomechanical model-based image registration method for aligning preoperative MR breast images with intraoperative patient positioning.
• To simultaneously optimize material parameters and gravitational force direction for accurate breast deformation modeling.

Main Methods:

• Utilized finite element (FE) based biomechanical modeling to simulate breast deformation under gravity.
• Integrated nonrigid intensity-based image registration to refine deformations not captured by FE models.
• Optimized material parameters and gravitational force direction concurrently.

Main Results:

• The proposed method achieved superior registration accuracy on skin fiducial markers (target registration error: 8.44 ±5.5 mm).
• Demonstrated higher overlap rates for fibroglandular tissue segmentation propagation (Dice Similarity Coefficient > 82%).
• Outperformed two alternative nonrigid registration methods in aligning prone and supine MR image pairs.

Conclusions:

• The proposed biomechanical model-based image registration effectively addresses the challenge of large breast deformations between imaging and surgical positions.
• This approach enhances the precision of surgical guidance, potentially leading to more complete cancer excision with adequate margins.
• The method shows significant promise for improving preoperative planning and intraoperative accuracy in breast conserving surgery.